Download A note on the variation of specific heats in ideal gases Most diatomic

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Heat wikipedia , lookup

Thermodynamic system wikipedia , lookup

First law of thermodynamics wikipedia , lookup

Heat capacity wikipedia , lookup

Equipartition theorem wikipedia , lookup

Adiabatic process wikipedia , lookup

Internal energy wikipedia , lookup

Conservation of energy wikipedia , lookup

T-symmetry wikipedia , lookup

History of thermodynamics wikipedia , lookup

Thermodynamic temperature wikipedia , lookup

Heat transfer physics wikipedia , lookup

Otto cycle wikipedia , lookup

Transcript
UT Austin ASE/EM Dept.
P. L. Varghese
A note on the variation of specific heats in ideal gases
Most diatomic gases such as nitrogen (N2) and oxygen (O2) at or near room temperature have
specific heats (cv and cp) that are almost constant. However, as the temperature (T) rises above
about 700 K, the specific heat begins to rise. Because the relation
c p (T ) − cv (T ) = R
remains valid, where R is the gas constant, it is sufficient to discuss (and calculate) the variation
of one of the specific heats with T. We shall talk about cv in this note.
In the following discussion all energies are mass specific, i.e. expressed per unit mass, and
denoted by lower case quantities. Energies per mole are obtained by multiplying mass specific
by the molar mass. In equations that contain the particular gas constant R, the molar energy is
obatined by substituting the universal gas constant R̂ for R.
Recall that "specific heat" is a somewhat misleading term that dates back to the time when heat
was viewed as a fluid that could be held by a material, hence terms like specific heat, heat
capacity, etc. Today we know that "heat" and "work" should always be understood as "heat
transfer" and "work transfer" because each is an energy transfer. Heat transfer is energy transfer
with entropy transfer; work transfer is energy transfer without entropy transfer. This is the only
way one can distinguish reliably between heat and work transfers.
The precise technical definitions of specific heat are
⎛ ∂u ⎞
⎛ ∂h ⎞
cv ≡ ⎜
⎟ ;cp ≡ ⎜
⎟ .
⎝ ∂T ⎠ v
⎝ ∂T ⎠ p
cv measures the rate of change of internal energy, u, with T. The internal energy u of any
substance is the energy stored in the molecules. In an ideal gas, the molecules are far apart on the
average, and the internal energy depends only on the energy stored in individual molecules, and
not on the separation between molecules. More precisely, the molecules are usually so far apart,
and spend so little time close to each other when they do collide, that the potential energy
associated with changing their average separation is negligible. If the energy did depend on the
average separation of molecules (i.e. the density), as it does in liquids and gases, then u = u(T,ρ),
whereas in an ideal gas u ≠ u(ρ). Thus, for ideal gases the partial derivatives above, become
ordinary derivatives:
du
dh
; cp =
cv =
.
dT
dT
For an ideal gas the internal energy is the sum of the kinetic energy of motion of the center of
mass of each individual molecule and the energy of motion relative to this center of mass. Even
if the bulk gas is at rest, the (centers of mass of) individual molecules are in constant chaotic
motion - that is sometimes referred to as the thermal energy. In fact, in non-equilibrium
situations, the temperature is often defined by the average kinetic energy of random motion of
the molecules, etr.
⎛1
⎞ 3
etr = 3 ⎜ RT ⎟ = RT
⎝2
⎠ 2
UT Austin ASE/EM Dept.
P. L. Varghese
For most diatomic molecules the rotation of the nuclei give the molecules an average rotational
energy erot that is given by
⎛1
⎞
erot = 2 ⎜ RT ⎟ = RT .
⎝2
⎠
(Hydrogen gas at low temperatures is an exception important in propulsion systems.) If we
assume that the internal energy is the sum of these two contributions
3
5
u = etr + erot = RT + RT = RT .
2
2
Then
cp 7
7
du 5
= R; c p = cv + R = R; γ =
= = 1.4
cv =
2
dT 2
cv 5
These results were first known from experiment, and were then explained by classical statistical
mechanics which predicted that at equilibrium each "degree of freedom" would contribute 1/2RT
to the internal energy. Thus translation has three degrees of freedom because the molecule can
move in 3-dimensional space, and a diatomic molecule has two additional degrees of freedom
because it can rotate about two orthogonal axes. (There is only a very small moment of inertia
about the internuclear axis, so no energy can be stored in rotation about this axis.) In this model
the nuclei are assumed to be effectively connected by a rigid massless rod, so the molecule
looked like a tiny dumbell.
However, classical statistical mechanics could not deal with the observed increase in specific
heat at high temperature. If it was assumed that the nuclei were not rigidly connected but could
vibrate, then one would have to admit two additional degrees of freedom to the system (potential
and kinetic energy of the oscillation) at all temperatures (not just high temperatures). In the
classical model
1
evib classical = 2 × RT = RT ;
2
7
u classical = etr + erot + evib classical = RT ;
2
7
9
9
cv ,classical = R; c p ,classical = R; γ classical = at all T !
2
7
2
Unfortunately, none of these predictions is in agreement with experiment, and the discovery that
there were free electrons in atoms made the situation worse, because now the energy of the
electrons in the molecule, eel, had to be accounted for too:
u = etr+ erot+ evib + eel.
The difficulties were only resolved with the introduction of quantum theory. Briefly, (and very
crudely) quantum theory explains that, just as matter comes in minimum sized packets called
atoms, energy and angular momentum come in packets too. Packets of energy are called quanta,
and the size of an energy quantum depends on the kind of energy you are dealing with. The
quanta of molecular translational energy are invariably so small that the "graininess" is not
apparent, and the translational energy of individual molecules can be treated as continuous, and
UT Austin ASE/EM Dept.
P. L. Varghese
we have "classical" behavior, i.e. negligible quantum effects. (In an analogous way the atomic
graininess of matter is usually not apparent when matter is treated in bulk and we use the
continuum approximation.)
In order to speak quantitatively it is convenient to use a characteristic temperature θ to describe
the size of a reference quantum of energy, ε = kθ, where k is Boltzmann's constant
(1.38×10-23J/K). Note that the quanta usually vary in size, and θ is used to describe a typical
value; the energy quanta are scaled to it depending on one or more quantum numbers. If T>>θ ,
then that type of energy can be well described by classical mechanics, but if T~θ then nonclassical effects are significant, and quantum mechanics must be used for an accurate
description. The following table gives typical values of θ for each different type of molecular
energy – the precise value depends on the properties of the atoms making up the molecule.
Energy type
Translation, θtr
Rotation, θrot
Vibration, θvib
Electronic, θel
Value of θ
<<10–10K (depends on molecular mass and size of the gas container; the larger
the containment volume the smaller the value of θ tr)
~0.5 – 2K (except 85K for H2)
~1000 – 3000K for most diatomics
~10,000 – 100,000K
The miniscule typical value of θtr emphasizes that quantum effects in translation are completely
negligible at ordinary temperatures, unless the molecule is confined in a container of atomic
dimensions like a very thin film. Hence etr= 3/2RT. We see that rotational quanta are also quite
small, so that at ordinary temperatures, rotation also behaves classically, i.e. erot=RT. (Hydrogen
gas is the common exception, and quantum mechanics explains why it is unusual.) However, the
vibrational quanta are usually large and electronic quanta are even bigger. Hence they behave
non-classically.
To predict the vibrational energy and the corresponding vibrational specific heat one needs a
model for the molecular vibration. If, as a simple approximation, one assumes that the molecule
vibrates like a simple harmonic oscillator, then quantum mechanics predicts that any single
molecule can have εvib = 0, kθvib, 2 kθvib, 3 kθvib, …but no other values. Statistical mechanics
predicts the precise probability that a randomly chosen molecule will have any one of these
energies. It also predicts that the average vibrational energy per unit mass of a collection of such
harmonic oscillators evib is given by:
Rθ vib
.
evib =
exp(θ vib / T ) − 1
Correspondingly,
⎡ (θ vib / 2T ) ⎤
de
= vib = R ⎢
⎥ .
dT
⎣ sinh (θ vib / 2T ) ⎦
2
cvib
UT Austin ASE/EM Dept.
P. L. Varghese
For temperatures up to about 3000 K, one can assume that the energy quanta required for
electronic excitation are so large that there is effectively no electronic energy. Thus eel = 0 at all
T of interest, and the corresponding electronic specific heat is also 0.
Finally, one can assemble all these results for the specific heats of a diatomic molecule:
du detr derot devib deel
=
+
+
+
= ctr + c rot + cvib + cel
cv =
dT
dT
dT
dT
dT
2
⎡ θ vib / 2T ⎤
⎡ θ vib / 2T ⎤
5
3
= R + R + R⎢
⎥ + 0 = R + R⎢
⎥
2
2
⎣ sinh (θ vib / 2T ) ⎦
⎣ sinh (θ vib / 2T ) ⎦
⎡ θ vib / 2T ⎤
7
c p = cv + R = R + R ⎢
⎥
2
⎣ sinh (θ vib / 2T ) ⎦
2
2
It is often convenient to write the specific heat as a sum of a low temperature limiting value and
the vibrational specific heat:
c p = c p _ low + cvib ;
7
R
2
These theoretical ideas were used to generate the model used to fit the data for the specific heat
of air. Because air is a mixture of nitrogen and oxygen, there are two different quantum
oscillators in air with different values of θvib. However, we only want an approximate fit to the
data, since in gas turbines the combustion products are not pure air anyway. Hence, a very
accurate model for cp for high temperature air is not justified for our cycle analyis. The data can
be satisfactorily fit to the equation:
c p _ low =
cp
R
=
c p _ low
R
2
c
⎡ θ /T ⎤
+ vib = 3.499 + α ⎢
⎥ ;
R
⎣ sinh (θ / T ) ⎦
α = 1.037; θ = θ vib / 2 = 1582 K
The quantity α is slightly different from unity to compensate for the fact that we are modelling
two different kinds of quantum oscillators as a single oscillator. Oxygen and nitrogen have
different values of θvib, and different molar masses (hence different R). This expression for cp(T)
is simple enough that it can be integrated analytically to give the following expressions for h(T)
and Δs:
⎡
2θ
⎛θ ⎞ ⎤
h(T ) = c p _ lowT + αθ ⎢coth⎜ ⎟ − 1⎥ = c p _ lowT + α
;
exp(2θ / T ) − 1
⎝T ⎠ ⎦
⎣
⎡⎛ θ ⎞
⎛ sinh(θ / T2 ) ⎞⎤
⎛θ ⎞
⎛θ ⎞ ⎛θ ⎞
p
Δs c p _ low T2
⎟⎟⎥
An − An 2 − α ⎢⎜⎜ ⎟⎟ coth⎜⎜ ⎟⎟ − ⎜⎜ ⎟⎟ coth ⎜⎜ ⎟⎟ + An⎜⎜
=
R
R
T1
p1
T
T
T
T
θ
T
sinh(
/
)
1 ⎠⎦
⎝
⎝ 2⎠
⎝ 1⎠ ⎝ 2⎠
⎣⎝ 1 ⎠
In each of these equations the term multiplying α gives the contribution of the vibrational
specific heat and the other term(s) give the perfect gas (constant specific heat) result.